feat: scoped-write interlock (roundtable-consensed, item #5 part 1) - #287
Open
CryptoJones wants to merge 1 commit into
Open
feat: scoped-write interlock (roundtable-consensed, item #5 part 1)#287CryptoJones wants to merge 1 commit into
CryptoJones wants to merge 1 commit into
Conversation
The scope field gains its second, ratified role: a guard-enforced operational interlock against accidental cross-project writes. Not a security boundary — deliberately bypassable — but it turns a fat-fingered out-of-scope write into a loud, remediable error instead of a silent one. Per the 2026-08-27 consensus: - Deny an out-of-scope write only when BOTH sides declare scope (process via OMIND_SCOPE, note via its Scope: field) and they differ. Undeclared OMIND_SCOPE fails open; an unscoped note is always writable (unscoped = global, the retrieval asymmetry). Escape hatch: OMIND_SCOPE_MODE=warn downgrades the deny to a returned scope_warning. Deny messages spell out the remediation. - Enforced on the agent-facing create-note / edit-note tools only, so mesh/system writes (which call the store directly) are exempt by design — the first merge never trips its own guard. create-note gains a scope param; edit guards the effective post-edit scope, and an edit that leaves scope untouched is still guarded against the note's existing scope. - Docs reframed on Note.scope: the interlock role and its bypassability are now explicit; "NOT A SECURITY BOUNDARY" survives. The scratch tier (item #5 part 2) and its TTL expiry land with `omind maintain` (item #3) next. 1,030 tests / ruff / mypy strict green. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbCqLsAJaoeCcyqnTs31Vk
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars. ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
CryptoJones
added a commit
that referenced
this pull request
Aug 28, 2026
Machine-local, auto-expiring scratch notes, marked by a `.scratch.md` filename suffix (the option chosen for its minimal mesh footprint). - create-note gains `scratch: bool`; a scratch note is written as `<title>.scratch.md`. The suffix IS the mark — no note-body field. - Still a top-level `*.md`, so search and listings find it like any note. - Machine-local by default: mesh gitignores `*.scratch.md`, so it is never committed or replicated (the "ephemeral + committed to permanent history is a contradiction" ruling). Opt-in replicate = create it as a normal note. - TTL expiry is an `omind maintain` step: scratch notes untouched for 7 days (from LAST MODIFICATION — HAL9000's clock refinement) are ARCHIVED, never deleted. Reports on a dry run; archives under --apply. Machine-local, so it runs outside --sync. Completes item #5 (the scoped-write interlock is #287). 1,029 tests / ruff / mypy strict green. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NbCqLsAJaoeCcyqnTs31Vk
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Item #5 (part 1) of the 2026-08-27 roadmap consensus: the
scopefield gainsits second, ratified role — a guard-enforced operational interlock against
accidental cross-project writes. Not a security boundary; deliberately
bypassable; it just turns a fat-fingered out-of-scope write into a loud,
remediable error instead of a silent one.
What it does
OMIND_SCOPE, note via itsScope:field) and they differ. UndeclaredOMIND_SCOPEfails open; an unscoped note is always writable (unscoped =global). Escape hatch:
OMIND_SCOPE_MODE=warndowngrades the deny to areturned
scope_warning. Deny messages spell out remediation.create-note/edit-notetools only, somesh/system writes (which call the store directly) are exempt by design — the
first merge never trips its own guard.
create-notegains ascopeparam;edit guards the effective post-edit scope, and an edit that leaves scope
untouched is still guarded against the note's existing scope.
Note.scope: the interlock role and its bypassability arenow explicit; "NOT A SECURITY BOUNDARY" survives.
Verification
src/omind/scope_guard.py+tests/test_scope_guard.py(7 unit tests) and4 server integration tests. 1,030 tests / ruff / mypy --strict green locally.
Part 2 (scratch tier) and item #3 (
omind maintain) follow.Note: repo CI has a pre-existing red state on
main(anmcpdependency-driftthat changes
ToolErrormessage text — 5test_servererror-message tests +Windows flakes). This PR does not touch it; its own new code is green.
Proudly Made in Nebraska. Go Big Red! 🌽 https://xkcd.com/2347/
🤖 Generated with Claude Code